OpenAI has launched a new artificial intelligence tool called “Deep Research,” designed to conduct multi-step research tasks on the internet and synthesize information into comprehensive reports.
The release follows the company’s recent debut of “Operator,” a tool that assists users with tasks such as grocery shopping and booking restaurant reservations.
Deep Research, unveiled during a demonstration in Washington and later on YouTube, is based on OpenAI’s upcoming o3 model, which is optimized for web browsing and data analysis. According to Kevin Weil, OpenAI’s Chief Product Officer, the tool can complete research tasks in five to 30 minutes, depending on complexity—work that would typically take a human from 30 minutes to 30 days.
Weil demonstrated the technology by generating a detailed report on Albert Einstein, including his background, personality, and suggested questions for a hypothetical Senate hearing.
“It can surf the web, understand text, images, and PDFs, and perform recursive searches to synthesize learned information,” Weil explained.
Deep Research is tailored for professionals in fields such as finance, science, law, and policy, where thorough and reliable research is crucial. It can also assist discerning consumers by providing personalized recommendations for complex purchasing decisions, like cars or appliances.
The tool generates reports with citations for information sources, aiming to increase transparency and accuracy. However, OpenAI acknowledged that Deep Research might sometimes struggle to differentiate authoritative information from rumors or convey uncertainty about data findings.
Despite its advanced capabilities, OpenAI warned that the tool is still in its early stages and prone to errors, including the phenomenon of “hallucination,” where AI systems produce incorrect information or citations.
The New York Times has filed a lawsuit against OpenAI and Microsoft, accusing them of copyright infringement for AI-generated content. OpenAI and Microsoft have denied the claims.
Deep Research became available on Sunday for ChatGPT Pro subscribers, a $200-a-month service providing access to OpenAI’s latest tools. The company plans to roll out the feature to other platforms, including desktop and mobile apps, by the end of February.
With input from Reuters, the New York Times, and OpenAI.
